1. Technical Field
The present invention relates to an image processing apparatus, an image processing method, a recording method, and a recording medium.
2. Related Art
There are conventionally known techniques for displaying a stereoscopic (3D) image. One technique for displaying a stereoscopic image involves generating a right-eye image of an object seen by the right eye and a left-eye image of the object seen by the left eye, and displaying the right-eye image to only the right eye of a user and the left-eye image to only the left eye of the user. This technique is referred to as stereoscopic viewing.    Patent Document 1: Japanese Patent Application Publication No. H10-117362    Patent Document 2: Japanese Patent Application Publication No. 2002-262310    Patent Document 3: Japanese Patent Application Publication No. H08-205201
This stereoscopic viewing technology, however, provides an image that appears unnatural to the user. With stereoscopic viewing technology, an object image displayed on the screen appears to jump out or fall back due to a parallax effect between the viewpoint of each eye of the user. Therefore, compared to looking at a normal stereoscopic object, the stereoscopic image is perceived differently by the brain, which can make the object appear unnatural. A main cause of this phenomenon is a mismatch between the convergence and the adjustment of sight, i.e. when the point at which the lines of sight intersect differs from the correct focal point of the actual object.
Furthermore, stereoscopic viewing technology is expensive, due to the need to generate parallax images to achieve the stereoscopic image. In stereoscopic viewing technology, it is necessary to generate two images, e.g. a left-side image and a right-side image, that create a parallax effect when viewed from a single location. Accordingly, when capturing an image, the parallax images must be captured by two or more cameras that are arranged and controlled appropriately, and this incurs a high cost. Furthermore, it is also expensive to generate parallax images when combining artificial images, such as in computer graphics.
There are various methods proposed for generating the parallax images based on a two-dimensional image. One such method involves displaying a plurality of images captured at different times to the right eye and the left eye, respectively. Another example involves generating the parallax images from the two-dimensional image by analyzing the two-dimensional image and having the user attach depth information thereto. However, all of these methods involve complex manual operations and judgments by the user, and also have high cost.
Another method proposed involves stereoscopically displaying an image without creating a parallax effect between the image for the left eye and the image for the right eye, as described in Patent Documents 1 to 3, for example. With these methods, however, there is no clear and specific process for suitably displaying the stereoscopic image. Furthermore, the stereoscopic image provided based on such a two-dimensional image includes portions at the right and left edges thereof in which only one of the images is displayed, and this causes the stereoscopic image to appear unnatural.
This stereoscopic viewing technology can also cause images at the edges of the display region to appear unnatural. For example, the portions of the stereoscopic image corresponding to the edges of the images can be cut off.